1# Changes/Release Notes 2 3On this page we provide a summary of the main API changes, new features and examples 4for each release of libCEED. 5 6(main)= 7 8## Current `main` branch 9 10### Interface changes 11 12- Update {c:func}`CeedQFunctionGetFields` and {c:func}`CeedOperatorGetFields` to include number of fields. 13- Promote to the public API: QFunction and Operator field objects, `CeedQFunctionField` and `CeedOperatorField`, and associated getters, {c:func}`CeedQFunctionGetFields`; {c:func}`CeedQFunctionFieldGetName`; {c:func}`CeedQFunctionFieldGetSize`; {c:func}`CeedQFunctionFieldGetEvalMode`; {c:func}`CeedOperatorGetFields`; {c:func}`CeedOperatorFieldGetElemRestriction`; {c:func}`CeedOperatorFieldGetBasis`; and {c:func}`CeedOperatorFieldGetVector`. 14- Clarify and document conditions where `CeedQFunction` and `CeedOperator` become immutable and no further fields or suboperators can be added. 15- Add {c:func}`CeedOperatorLinearAssembleQFunctionBuildOrUpdate` to reduce object creation overhead in assembly of CeedOperator preconditioning ingredients. 16- Promote {c:func}`CeedOperatorCheckReady`to the public API to facilitate interactive interfaces. 17- Warning added when compiling OCCA backend to alert users that this backend is experimental. 18- `ceed-backend.h`, `ceed-hash.h`, and `ceed-khash.h` removed. Users should use `ceed/backend.h`, `ceed/hash.h`, and `ceed/khash.h`. 19- Added {c:func}`CeedQFunctionGetKernelName`; refactored {c:func}`CeedQFunctionGetSourcePath` to exclude function kernel name. 20- Clarify documentation for {c:func}`CeedVectorTakeArray`; this function will error if {c:func}`CeedVectorSetArray` with `copy_mode == CEED_USE_POINTER` was not previously called for the corresponding `CeedMemType`. 21- Added {c:func}`CeedVectorGetArrayWrite` that allows access to uninitalized arrays; require initalized data for {c:func}`CeedVectorGetArray`. 22 23### New features 24 25- `CeedScalar` can now be set as `float` or `double` at compile time. 26- Added JiT utilities in `ceed/jit-tools.h` to reduce duplicated code in GPU backends. 27- Added support for JiT of QFunctions with `#include "relative/path/local-file.h"` statements for additional local files. Note that files included with `""` are searched relative to the current file first, then by compiler paths (as with `<>` includes). To use this feature, one should adhere to relative paths only, not compiler flags like `-I`, which the JiT will not be aware of. 28- Remove need to guard library headers in QFunction source for code generation backends. 29- `CeedDebugEnv()` macro created to provide debugging outputs when Ceed context is not present. 30 31### Maintainability 32 33- Refactored preconditioner support internally to facilitate future development and improve GPU completeness/test coverage. 34- `Include-what-you-use` makefile target added as `make iwyu`. 35- Create backend constant `CEED_FIELD_MAX` to reduce magic numbers in codebase. 36 37(v0-9)= 38 39## v0.9 (Jul 6, 2021) 40 41### Interface changes 42 43- Minor modification in error handling macro to silence pedantic warnings when compiling with Clang, but no functional impact. 44 45### New features 46 47- Add {c:func}`CeedVectorAXPY` and {c:func}`CeedVectorPointwiseMult` as a convenience for stand-alone testing and internal use. 48- Add `CEED_QFUNCTION_HELPER` macro to properly annotate QFunction helper functions for code generation backends. 49- Add `CeedPragmaOptimizeOff` macro for code that is sensitive to floating point errors from fast math optimizations. 50- Rust support: split `libceed-sys` crate out of `libceed` and [publish both on crates.io](https://crates.io/crates/libceed). 51 52### Performance improvements 53 54### Examples 55 56- Solid mechanics mini-app updated to explore the performance impacts of various formulations in the initial and current configurations. 57- Fluid mechanics example adds GPU support and improves modularity. 58 59### Deprecated backends 60 61- The `/cpu/self/tmpl` and `/cpu/self/tmpl/sub` backends have been removed. These backends were intially added to test the backend inheritance mechanism, but this mechanism is now widely used and tested in multiple backends. 62 63(v0-8)= 64 65## v0.8 (Mar 31, 2021) 66 67### Interface changes 68 69- Error handling improved to include enumerated error codes for C interface return values. 70- Installed headers that will follow semantic versioning were moved to {code}`include/ceed` directory. These headers have been renamed from {code}`ceed-*.h` to {code}`ceed/*.h`. Placeholder headers with the old naming schema are currently provided, but these headers will be removed in the libCEED v0.9 release. 71 72### New features 73 74- Julia and Rust interfaces added, providing a nearly 1-1 correspondence with the C interface, plus some convenience features. 75- Static libraries can be built with `make STATIC=1` and the pkg-config file is installed accordingly. 76- Add {c:func}`CeedOperatorLinearAssembleSymbolic` and {c:func}`CeedOperatorLinearAssemble` to support full assembly of libCEED operators. 77 78### Performance improvements 79 80- New HIP MAGMA backends for hipMAGMA library users: `/gpu/hip/magma` and `/gpu/hip/magma/det`. 81- New HIP backends for improved tensor basis performance: `/gpu/hip/shared` and `/gpu/hip/gen`. 82 83### Examples 84 85- {ref}`example-petsc-elasticity` example updated with traction boundary conditions and improved Dirichlet boundary conditions. 86- {ref}`example-petsc-elasticity` example updated with Neo-Hookean hyperelasticity in current configuration as well as improved Neo-Hookean hyperelasticity exploring storage vs computation tradeoffs. 87- {ref}`example-petsc-navier-stokes` example updated with isentropic traveling vortex test case, an analytical solution to the Euler equations that is useful for testing boundary conditions, discretization stability, and order of accuracy. 88- {ref}`example-petsc-navier-stokes` example updated with support for performing convergence study and plotting order of convergence by polynomial degree. 89 90(v0-7)= 91 92## v0.7 (Sep 29, 2020) 93 94### Interface changes 95 96- Replace limited {code}`CeedInterlaceMode` with more flexible component stride {code}`compstride` in {code}`CeedElemRestriction` constructors. 97 As a result, the {code}`indices` parameter has been replaced with {code}`offsets` and the {code}`nnodes` parameter has been replaced with {code}`lsize`. 98 These changes improve support for mixed finite element methods. 99- Replace various uses of {code}`Ceed*Get*Status` with {code}`Ceed*Is*` in the backend API to match common nomenclature. 100- Replace {code}`CeedOperatorAssembleLinearDiagonal` with {c:func}`CeedOperatorLinearAssembleDiagonal` for clarity. 101- Linear Operators can be assembled as point-block diagonal matrices with {c:func}`CeedOperatorLinearAssemblePointBlockDiagonal`, provided in row-major form in a {code}`ncomp` by {code}`ncomp` block per node. 102- Diagonal assemble interface changed to accept a {ref}`CeedVector` instead of a pointer to a {ref}`CeedVector` to reduce memory movement when interfacing with calling code. 103- Added {c:func}`CeedOperatorLinearAssembleAddDiagonal` and {c:func}`CeedOperatorLinearAssembleAddPointBlockDiagonal` for improved future integration with codes such as MFEM that compose the action of {ref}`CeedOperator`s external to libCEED. 104- Added {c:func}`CeedVectorTakeAray` to sync and remove libCEED read/write access to an allocated array and pass ownership of the array to the caller. 105 This function is recommended over {c:func}`CeedVectorSyncArray` when the {code}`CeedVector` has an array owned by the caller that was set by {c:func}`CeedVectorSetArray`. 106- Added {code}`CeedQFunctionContext` object to manage user QFunction context data and reduce copies between device and host memory. 107- Added {c:func}`CeedOperatorMultigridLevelCreate`, {c:func}`CeedOperatorMultigridLevelCreateTensorH1`, and {c:func}`CeedOperatorMultigridLevelCreateH1` to facilitate creation of multigrid prolongation, restriction, and coarse grid operators using a common quadrature space. 108 109### New features 110 111- New HIP backend: `/gpu/hip/ref`. 112- CeedQFunction support for user `CUfunction`s in some backends 113 114### Performance improvements 115 116- OCCA backend rebuilt to facilitate future performance enhancements. 117- Petsc BPs suite improved to reduce noise due to multiple calls to {code}`mpiexec`. 118 119### Examples 120 121- {ref}`example-petsc-elasticity` example updated with strain energy computation and more flexible boundary conditions. 122 123### Deprecated backends 124 125- The `/gpu/cuda/reg` backend has been removed, with its core features moved into `/gpu/cuda/ref` and `/gpu/cuda/shared`. 126 127(v0-6)= 128 129## v0.6 (Mar 29, 2020) 130 131libCEED v0.6 contains numerous new features and examples, as well as expanded 132documentation in [this new website](https://libceed.readthedocs.io). 133 134### New features 135 136- New Python interface using [CFFI](https://cffi.readthedocs.io/) provides a nearly 137 1-1 correspondence with the C interface, plus some convenience features. For instance, 138 data stored in the {cpp:type}`CeedVector` structure are available without copy as 139 {py:class}`numpy.ndarray`. Short tutorials are provided in 140 [Binder](https://mybinder.org/v2/gh/CEED/libCEED/main?urlpath=lab/tree/examples/tutorials/). 141- Linear QFunctions can be assembled as block-diagonal matrices (per quadrature point, 142 {c:func}`CeedOperatorAssembleLinearQFunction`) or to evaluate the diagonal 143 ({c:func}`CeedOperatorAssembleLinearDiagonal`). These operations are useful for 144 preconditioning ingredients and are used in the libCEED's multigrid examples. 145- The inverse of separable operators can be obtained using 146 {c:func}`CeedOperatorCreateFDMElementInverse` and applied with 147 {c:func}`CeedOperatorApply`. This is a useful preconditioning ingredient, 148 especially for Laplacians and related operators. 149- New functions: {c:func}`CeedVectorNorm`, {c:func}`CeedOperatorApplyAdd`, 150 {c:func}`CeedQFunctionView`, {c:func}`CeedOperatorView`. 151- Make public accessors for various attributes to facilitate writing composable code. 152- New backend: `/cpu/self/memcheck/serial`. 153- QFunctions using variable-length array (VLA) pointer constructs can be used with CUDA 154 backends. (Single source is coming soon for OCCA backends.) 155- Fix some missing edge cases in CUDA backend. 156 157### Performance Improvements 158 159- MAGMA backend performance optimization and non-tensor bases. 160- No-copy optimization in {c:func}`CeedOperatorApply`. 161 162### Interface changes 163 164- Replace {code}`CeedElemRestrictionCreateIdentity` and 165 {code}`CeedElemRestrictionCreateBlocked` with more flexible 166 {c:func}`CeedElemRestrictionCreateStrided` and 167 {c:func}`CeedElemRestrictionCreateBlockedStrided`. 168- Add arguments to {c:func}`CeedQFunctionCreateIdentity`. 169- Replace ambiguous uses of {cpp:enum}`CeedTransposeMode` for L-vector identification 170 with {cpp:enum}`CeedInterlaceMode`. This is now an attribute of the 171 {cpp:type}`CeedElemRestriction` (see {c:func}`CeedElemRestrictionCreate`) and no 172 longer passed as `lmode` arguments to {c:func}`CeedOperatorSetField` and 173 {c:func}`CeedElemRestrictionApply`. 174 175### Examples 176 177libCEED-0.6 contains greatly expanded examples with {ref}`new documentation <Examples>`. 178Notable additions include: 179 180- Standalone {ref}`ex2-surface` ({file}`examples/ceed/ex2-surface`): compute the area of 181 a domain in 1, 2, and 3 dimensions by applying a Laplacian. 182 183- PETSc {ref}`example-petsc-area` ({file}`examples/petsc/area.c`): computes surface area 184 of domains (like the cube and sphere) by direct integration on a surface mesh; 185 demonstrates geometric dimension different from topological dimension. 186 187- PETSc {ref}`example-petsc-bps`: 188 189 - {file}`examples/petsc/bpsraw.c` (formerly `bps.c`): transparent CUDA support. 190 - {file}`examples/petsc/bps.c` (formerly `bpsdmplex.c`): performance improvements 191 and transparent CUDA support. 192 - {ref}`example-petsc-bps-sphere` ({file}`examples/petsc/bpssphere.c`): 193 generalizations of all CEED BPs to the surface of the sphere; demonstrates geometric 194 dimension different from topological dimension. 195 196- {ref}`example-petsc-multigrid` ({file}`examples/petsc/multigrid.c`): new p-multigrid 197 solver with algebraic multigrid coarse solve. 198 199- {ref}`example-petsc-navier-stokes` ({file}`examples/fluids/navierstokes.c`; formerly 200 `examples/navier-stokes`): unstructured grid support (using PETSc's `DMPlex`), 201 implicit time integration, SU/SUPG stabilization, free-slip boundary conditions, and 202 quasi-2D computational domain support. 203 204- {ref}`example-petsc-elasticity` ({file}`examples/solids/elasticity.c`): new solver for 205 linear elasticity, small-strain hyperelasticity, and globalized finite-strain 206 hyperelasticity using p-multigrid with algebraic multigrid coarse solve. 207 208(v0-5)= 209 210## v0.5 (Sep 18, 2019) 211 212For this release, several improvements were made. Two new CUDA backends were added to 213the family of backends, of which, the new `cuda-gen` backend achieves state-of-the-art 214performance using single-source {ref}`CeedQFunction`. From this release, users 215can define Q-Functions in a single source code independently of the targeted backend 216with the aid of a new macro `CEED QFUNCTION` to support JIT (Just-In-Time) and CPU 217compilation of the user provided {ref}`CeedQFunction` code. To allow a unified 218declaration, the {ref}`CeedQFunction` API has undergone a slight change: 219the `QFunctionField` parameter `ncomp` has been changed to `size`. This change 220requires setting the previous value of `ncomp` to `ncomp*dim` when adding a 221`QFunctionField` with eval mode `CEED EVAL GRAD`. 222 223Additionally, new CPU backends 224were included in this release, such as the `/cpu/self/opt/*` backends (which are 225written in pure C and use partial **E-vectors** to improve performance) and the 226`/cpu/self/ref/memcheck` backend (which relies upon the 227[Valgrind](http://valgrind.org/) Memcheck tool to help verify that user 228{ref}`CeedQFunction` have no undefined values). 229This release also included various performance improvements, bug fixes, new examples, 230and improved tests. Among these improvements, vectorized instructions for 231{ref}`CeedQFunction` code compiled for CPU were enhanced by using `CeedPragmaSIMD` 232instead of `CeedPragmaOMP`, implementation of a {ref}`CeedQFunction` gallery and 233identity Q-Functions were introduced, and the PETSc benchmark problems were expanded 234to include unstructured meshes handling were. For this expansion, the prior version of 235the PETSc BPs, which only included data associated with structured geometries, were 236renamed `bpsraw`, and the new version of the BPs, which can handle data associated 237with any unstructured geometry, were called `bps`. Additionally, other benchmark 238problems, namely BP2 and BP4 (the vector-valued versions of BP1 and BP3, respectively), 239and BP5 and BP6 (the collocated versions---for which the quadrature points are the same 240as the Gauss Lobatto nodes---of BP3 and BP4 respectively) were added to the PETSc 241examples. Furthermoew, another standalone libCEED example, called `ex2`, which 242computes the surface area of a given mesh was added to this release. 243 244Backends available in this release: 245 246| CEED resource (`-ceed`) | Backend | 247|--------------------------|-----------------------------------------------------| 248| `/cpu/self/ref/serial` | Serial reference implementation | 249| `/cpu/self/ref/blocked` | Blocked reference implementation | 250| `/cpu/self/ref/memcheck` | Memcheck backend, undefined value checks | 251| `/cpu/self/opt/serial` | Serial optimized C implementation | 252| `/cpu/self/opt/blocked` | Blocked optimized C implementation | 253| `/cpu/self/avx/serial` | Serial AVX implementation | 254| `/cpu/self/avx/blocked` | Blocked AVX implementation | 255| `/cpu/self/xsmm/serial` | Serial LIBXSMM implementation | 256| `/cpu/self/xsmm/blocked` | Blocked LIBXSMM implementation | 257| `/cpu/occa` | Serial OCCA kernels | 258| `/gpu/occa` | CUDA OCCA kernels | 259| `/omp/occa` | OpenMP OCCA kernels | 260| `/ocl/occa` | OpenCL OCCA kernels | 261| `/gpu/cuda/ref` | Reference pure CUDA kernels | 262| `/gpu/cuda/reg` | Pure CUDA kernels using one thread per element | 263| `/gpu/cuda/shared` | Optimized pure CUDA kernels using shared memory | 264| `/gpu/cuda/gen` | Optimized pure CUDA kernels using code generation | 265| `/gpu/magma` | CUDA MAGMA kernels | 266 267Examples available in this release: 268 269:::{list-table} 270:header-rows: 1 271:widths: auto 272* - User code 273 - Example 274* - `ceed` 275 - * ex1 (volume) 276 * ex2 (surface) 277* - `mfem` 278 - * BP1 (scalar mass operator) 279 * BP3 (scalar Laplace operator) 280* - `petsc` 281 - * BP1 (scalar mass operator) 282 * BP2 (vector mass operator) 283 * BP3 (scalar Laplace operator) 284 * BP4 (vector Laplace operator) 285 * BP5 (collocated scalar Laplace operator) 286 * BP6 (collocated vector Laplace operator) 287 * Navier-Stokes 288* - `nek5000` 289 - * BP1 (scalar mass operator) 290 * BP3 (scalar Laplace operator) 291::: 292 293(v0-4)= 294 295## v0.4 (Apr 1, 2019) 296 297libCEED v0.4 was made again publicly available in the second full CEED software 298distribution, release CEED 2.0. This release contained notable features, such as 299four new CPU backends, two new GPU backends, CPU backend optimizations, initial 300support for operator composition, performance benchmarking, and a Navier-Stokes demo. 301The new CPU backends in this release came in two families. The `/cpu/self/*/serial` 302backends process one element at a time and are intended for meshes with a smaller number 303of high order elements. The `/cpu/self/*/blocked` backends process blocked batches of 304eight interlaced elements and are intended for meshes with higher numbers of elements. 305The `/cpu/self/avx/*` backends rely upon AVX instructions to provide vectorized CPU 306performance. The `/cpu/self/xsmm/*` backends rely upon the 307[LIBXSMM](http://github.com/hfp/libxsmm) package to provide vectorized CPU 308performance. The `/gpu/cuda/*` backends provide GPU performance strictly using CUDA. 309The `/gpu/cuda/ref` backend is a reference CUDA backend, providing reasonable 310performance for most problem configurations. The `/gpu/cuda/reg` backend uses a simple 311parallelization approach, where each thread treats a finite element. Using just in time 312compilation, provided by nvrtc (NVidia Runtime Compiler), and runtime parameters, this 313backend unroll loops and map memory address to registers. The `/gpu/cuda/reg` backend 314achieve good peak performance for 1D, 2D, and low order 3D problems, but performance 315deteriorates very quickly when threads run out of registers. 316 317A new explicit time-stepping Navier-Stokes solver was added to the family of libCEED 318examples in the `examples/petsc` directory (see {ref}`example-petsc-navier-stokes`). 319This example solves the time-dependent Navier-Stokes equations of compressible gas 320dynamics in a static Eulerian three-dimensional frame, using structured high-order 321finite/spectral element spatial discretizations and explicit high-order time-stepping 322(available in PETSc). Moreover, the Navier-Stokes example was developed using PETSc, 323so that the pointwise physics (defined at quadrature points) is separated from the 324parallelization and meshing concerns. 325 326Backends available in this release: 327 328| CEED resource (`-ceed`) | Backend | 329|--------------------------|-----------------------------------------------------| 330| `/cpu/self/ref/serial` | Serial reference implementation | 331| `/cpu/self/ref/blocked` | Blocked reference implementation | 332| `/cpu/self/tmpl` | Backend template, defaults to `/cpu/self/blocked` | 333| `/cpu/self/avx/serial` | Serial AVX implementation | 334| `/cpu/self/avx/blocked` | Blocked AVX implementation | 335| `/cpu/self/xsmm/serial` | Serial LIBXSMM implementation | 336| `/cpu/self/xsmm/blocked` | Blocked LIBXSMM implementation | 337| `/cpu/occa` | Serial OCCA kernels | 338| `/gpu/occa` | CUDA OCCA kernels | 339| `/omp/occa` | OpenMP OCCA kernels | 340| `/ocl/occa` | OpenCL OCCA kernels | 341| `/gpu/cuda/ref` | Reference pure CUDA kernels | 342| `/gpu/cuda/reg` | Pure CUDA kernels using one thread per element | 343| `/gpu/magma` | CUDA MAGMA kernels | 344 345Examples available in this release: 346 347:::{list-table} 348:header-rows: 1 349:widths: auto 350* - User code 351 - Example 352* - `ceed` 353 - * ex1 (volume) 354* - `mfem` 355 - * BP1 (scalar mass operator) 356 * BP3 (scalar Laplace operator) 357* - `petsc` 358 - * BP1 (scalar mass operator) 359 * BP3 (scalar Laplace operator) 360 * Navier-Stokes 361* - `nek5000` 362 - * BP1 (scalar mass operator) 363 * BP3 (scalar Laplace operator) 364::: 365 366(v0-3)= 367 368## v0.3 (Sep 30, 2018) 369 370Notable features in this release include active/passive field interface, support for 371non-tensor bases, backend optimization, and improved Fortran interface. This release 372also focused on providing improved continuous integration, and many new tests with code 373coverage reports of about 90%. This release also provided a significant change to the 374public interface: a {ref}`CeedQFunction` can take any number of named input and output 375arguments while {ref}`CeedOperator` connects them to the actual data, which may be 376supplied explicitly to `CeedOperatorApply()` (active) or separately via 377`CeedOperatorSetField()` (passive). This interface change enables reusable libraries 378of CeedQFunctions and composition of block solvers constructed using 379{ref}`CeedOperator`. A concept of blocked restriction was added to this release and 380used in an optimized CPU backend. Although this is typically not visible to the user, 381it enables effective use of arbitrary-length SIMD while maintaining cache locality. 382This CPU backend also implements an algebraic factorization of tensor product gradients 383to perform fewer operations than standard application of interpolation and 384differentiation from nodes to quadrature points. This algebraic formulation 385automatically supports non-polynomial and non-interpolatory bases, thus is more general 386than the more common derivation in terms of Lagrange polynomials on the quadrature points. 387 388Backends available in this release: 389 390| CEED resource (`-ceed`) | Backend | 391|-------------------------|-----------------------------------------------------| 392| `/cpu/self/blocked` | Blocked reference implementation | 393| `/cpu/self/ref` | Serial reference implementation | 394| `/cpu/self/tmpl` | Backend template, defaults to `/cpu/self/blocked` | 395| `/cpu/occa` | Serial OCCA kernels | 396| `/gpu/occa` | CUDA OCCA kernels | 397| `/omp/occa` | OpenMP OCCA kernels | 398| `/ocl/occa` | OpenCL OCCA kernels | 399| `/gpu/magma` | CUDA MAGMA kernels | 400 401Examples available in this release: 402 403:::{list-table} 404:header-rows: 1 405:widths: auto 406* - User code 407 - Example 408* - `ceed` 409 - * ex1 (volume) 410* - `mfem` 411 - * BP1 (scalar mass operator) 412 * BP3 (scalar Laplace operator) 413* - `petsc` 414 - * BP1 (scalar mass operator) 415 * BP3 (scalar Laplace operator) 416* - `nek5000` 417 - * BP1 (scalar mass operator) 418 * BP3 (scalar Laplace operator) 419::: 420 421(v0-21)= 422 423## v0.21 (Sep 30, 2018) 424 425A MAGMA backend (which relies upon the 426[MAGMA](https://bitbucket.org/icl/magma) package) was integrated in libCEED for this 427release. This initial integration set up the framework of using MAGMA and provided the 428libCEED functionality through MAGMA kernels as one of libCEED’s computational backends. 429As any other backend, the MAGMA backend provides extended basic data structures for 430{ref}`CeedVector`, {ref}`CeedElemRestriction`, and {ref}`CeedOperator`, and implements 431the fundamental CEED building blocks to work with the new data structures. 432In general, the MAGMA-specific data structures keep the libCEED pointers to CPU data 433but also add corresponding device (e.g., GPU) pointers to the data. Coherency is handled 434internally, and thus seamlessly to the user, through the functions/methods that are 435provided to support them. 436 437Backends available in this release: 438 439| CEED resource (`-ceed`) | Backend | 440|-------------------------|---------------------------------| 441| `/cpu/self` | Serial reference implementation | 442| `/cpu/occa` | Serial OCCA kernels | 443| `/gpu/occa` | CUDA OCCA kernels | 444| `/omp/occa` | OpenMP OCCA kernels | 445| `/ocl/occa` | OpenCL OCCA kernels | 446| `/gpu/magma` | CUDA MAGMA kernels | 447 448Examples available in this release: 449 450:::{list-table} 451:header-rows: 1 452:widths: auto 453* - User code 454 - Example 455* - `ceed` 456 - * ex1 (volume) 457* - `mfem` 458 - * BP1 (scalar mass operator) 459 * BP3 (scalar Laplace operator) 460* - `petsc` 461 - * BP1 (scalar mass operator) 462* - `nek5000` 463 - * BP1 (scalar mass operator) 464::: 465 466(v0-2)= 467 468## v0.2 (Mar 30, 2018) 469 470libCEED was made publicly available the first full CEED software distribution, release 471CEED 1.0. The distribution was made available using the Spack package manager to provide 472a common, easy-to-use build environment, where the user can build the CEED distribution 473with all dependencies. This release included a new Fortran interface for the library. 474This release also contained major improvements in the OCCA backend (including a new 475`/ocl/occa` backend) and new examples. The standalone libCEED example was modified to 476compute the volume volume of a given mesh (in 1D, 2D, or 3D) and placed in an 477`examples/ceed` subfolder. A new `mfem` example to perform BP3 (with the application 478of the Laplace operator) was also added to this release. 479 480Backends available in this release: 481 482| CEED resource (`-ceed`) | Backend | 483|-------------------------|---------------------------------| 484| `/cpu/self` | Serial reference implementation | 485| `/cpu/occa` | Serial OCCA kernels | 486| `/gpu/occa` | CUDA OCCA kernels | 487| `/omp/occa` | OpenMP OCCA kernels | 488| `/ocl/occa` | OpenCL OCCA kernels | 489 490Examples available in this release: 491 492:::{list-table} 493:header-rows: 1 494:widths: auto 495* - User code 496 - Example 497* - `ceed` 498 - * ex1 (volume) 499* - `mfem` 500 - * BP1 (scalar mass operator) 501 * BP3 (scalar Laplace operator) 502* - `petsc` 503 - * BP1 (scalar mass operator) 504* - `nek5000` 505 - * BP1 (scalar mass operator) 506::: 507 508(v0-1)= 509 510## v0.1 (Jan 3, 2018) 511 512Initial low-level API of the CEED project. The low-level API provides a set of Finite 513Elements kernels and components for writing new low-level kernels. Examples include: 514vector and sparse linear algebra, element matrix assembly over a batch of elements, 515partial assembly and action for efficient high-order operators like mass, diffusion, 516advection, etc. The main goal of the low-level API is to establish the basis for the 517high-level API. Also, identifying such low-level kernels and providing a reference 518implementation for them serves as the basis for specialized backend implementations. 519This release contained several backends: `/cpu/self`, and backends which rely upon the 520[OCCA](http://github.com/libocca/occa) package, such as `/cpu/occa`, 521`/gpu/occa`, and `/omp/occa`. 522It also included several examples, in the `examples` folder: 523A standalone code that shows the usage of libCEED (with no external 524dependencies) to apply the Laplace operator, `ex1`; an `mfem` example to perform BP1 525(with the application of the mass operator); and a `petsc` example to perform BP1 526(with the application of the mass operator). 527 528Backends available in this release: 529 530| CEED resource (`-ceed`) | Backend | 531|-------------------------|---------------------------------| 532| `/cpu/self` | Serial reference implementation | 533| `/cpu/occa` | Serial OCCA kernels | 534| `/gpu/occa` | CUDA OCCA kernels | 535| `/omp/occa` | OpenMP OCCA kernels | 536 537Examples available in this release: 538 539| User code | Example | 540|-----------------------|-----------------------------------| 541| `ceed` | ex1 (scalar Laplace operator) | 542| `mfem` | BP1 (scalar mass operator) | 543| `petsc` | BP1 (scalar mass operator) | 544``` 545